Role of HIF-1α-miR30a-Snai1 Axis in Neonatal Hyperoxic Lung Injury

Bronchopulmonary dysplasia (BPD) is characterized by a severe impairment in lung alveolarization and vascular development. We have previously shown that pulmonary angiogenesis is preserved in hyperoxia-exposed female mice accompanied by increased miR-30a expression, which is a proangiogenic miRNA. Also, miR-30a expression is decreased in human BPD. HIF-1α plays an essential role in postnatal lung development, especially in recovery from hyperoxic injury. Snai1 activation promotes pathological fibrosis through many mechanisms including Endo-MT, which may in turn adversely impact lung vascular development. Our objective was to test the hypothesis that higher miR-30a expression through HIF-1α decreases Snai1 expression in females and attenuates injury in the developing lung. Neonatal male and female mice (C57BL/6) were exposed to hyperoxia (P1-5, 0.95 FiO2) and euthanized on P21. Neonatal human pulmonary microvascular endothelial cells (HPMECs; 18-24-week gestation donors; 3/group either sex) were subjected to hyperoxia (95% O2 and 5% CO2) or normoxia (air and 5% CO2) up to 72 h. Snai1 expression was measured in HPMECs in vitro and in neonatal mouse lungs in vivo. Also, Snai1 expression was measured in HPMECs after miR-30a mimic and miR-30a inhibitor treatment. To further establish the potential regulation of miR-30a by Hif-1α, miR-30a expression after Hif-1α inhibition was measured in HPMECs. In vivo, Snai1 expression was decreased in neonatal female lungs compared to males at P7. Increased Snai1 expression was seen in male HPMECs upon exposure to hyperoxia in vitro. Treatment with the miR-30a mimic decreased Snai1 expression in HPMECs, while miR-30a inhibition significantly increased Snai1 expression in HPMECs. siRNA-mediated loss of Hif-1α expression in HPMECs decreased miR-30a expression. Hif-1α may lead to differential sex-specific miR-30a expression and may contribute to protection from hyperoxic lung injury in female neonatal mice through decreased Snai1 expression.


Introduction
With the increasing survival of extremely premature babies, the incidence of bronchopulmonary dysplasia (BPD) has remained steady, despite the advances in neonatal intensive care. BPD is the cause of significant morbidity and leads to prolonged impairment in lung function in this group of patients. Many prenatal and postnatal factors are involved in the pathogenesis of BPD, and exposure to supraphysiological concentrations of oxygen contributes to its development. Murine models have utilized varying durations of postnatal hyperoxia exposure to simulate the human disease in neonatal mice [1,2].
However, the mechanistic role of the intersection between HIF-1α, miR-30a, and Snai1 in neonatal hyperoxic lung injury has not been studied. Our objective was to test the hypothesis that HIF-1α increases the miR-30a expression in females and decreases the Snai1 expression in the developing lung.

Animals.
All experiments were performed per relevant guidelines and regulations. The care of animals was as per the 8 th edition of the guide for the care and use of laboratory animals and other IACUC protocols. Timed pregnant C57BL/6J WT mice were obtained from Charles River Laboratories (Wilmington). The sex in neonatal mouse pups was determined as described before and with PCR analysis for the Sry gene [8].
2.1.1. Mouse Model of BPD. Mouse pups were exposed to normoxia (21% O 2 ) or hyperoxia (95% O 2 ), within 12 h of birth for five days. Neonatal mice are at the saccular stage of lung development during this period, equivalent to 26-36 weeks in human gestation. The dams were rotated between air-and hyperoxia-exposed litters every 24 hours to prevent oxygen toxicity in the dams. Oxygen exposure was conducted in plexiglass chambers as previously described [8]. Mice were euthanized on P7 and P21 (after recovery in room air). The mice were euthanized with sodium pentobarbital, 100 mg/kg, i.p. For the in vivo experiments, neonatal mice from different litters were randomly allocated to room air or hyperoxia. The experimental unit was a single neonatal mouse pup and data from 4-6 animals/group were used for analysis.

Cell
Culture and Hyperoxia Treatment. Neonatal human pulmonary microvascular endothelial cells (HPMEC; male: #10885, #11367, and #10899; female: #5016, #10160, and #10169; 6 biological replicates in total) were purchased from ScienCell and maintained in Endothelial Cell Medium (Cat # 1001, ScienCell) at 37°C in 5% CO 2 . The gestational age of the donors varied from 18 to 24 weeks. Male and female HPMECs were used from passages 3 to 5 to ensure their endothelial characteristics. Male or female HPMEC cells (1 × 10 5 ) were seeded in a 6 mm dish. 24 h later, these cells were incubated at 37°C in room air condition (21% O 2 , 5% CO 2 ) or in hyperoxia (95% O 2 , 5% CO 2 ) as described before for 72 h. For measuring SNAI1 expression, both male and female HPMECs were used. For the rest of the in vitro experiments, female HPMECs were used as the miR-30a expression was increased only in female HPMECs [9].
2.1.6. Statistical Analysis. GraphPad version 7 was used for the analysis of our data. Data are expressed as means ± SD.
Data were analyzed by two-way ANOVA to test for the independent effects of sex and hyperoxia and to look for any interaction (sex × hyperoxia) or by Mann-Whitney U test. Multiple-comparison testing (Bonferroni) was performed if statistical significance (P < 0:05) was noted by ANOVA.

Differential Sex-Specific Expression of Snai1 in Neonatal
Mice after Postnatal Hyperoxia Exposure. WT male and female neonatal mice were exposed to hyperoxia (0.95 FiO 2 from P1-5) during the saccular stage of lung development, and Snai1 mRNA and protein expression were measured at P7 (early) and P21 (during recovery in normoxia) and compared to respective normoxic controls. Snai1 mRNA expression (β-actin as housekeeping gene) was decreased in female neonatal mice at P7 after hyperoxia exposure compared to female normoxic controls. No difference in expression was seen in males (Figure 1(a)). The results were similar with another housekeeping gene (β2-microglobulin) as shown in Supp. Figure 1. Similarly, Snai1 Values are means ± SD. n = 5 − 6 mice per group. Significant differences between groups indicated by * P < 0:05 and * * P < 0:01.
protein expression in the lungs was also decreased in females compared to female normoxic controls and compared to hyperoxia-exposed male mice (Figure 1(b)). Snai1 mRNA and protein expression did not show statistically different expression levels between any of the groups at P21 (Figures 1(c) and 1(d)). Protein expression was repeated with vinculin as the loading control; female mice tended to have lower Snai1 protein in the lungs at P7 compared to males, and in HPMECs, males tended to have a greater increase in Snai1 expression. These results however, were not statistically significant (Supp. Figure 1).

Snai1 mRNA and Protein Expression Are Increased in
Male Human Pulmonary Microvascular Endothelial Cells upon Exposure to Hyperoxia. Next, we wanted to elucidate the response to hyperoxia in neonatal human pulmonary microvascular endothelial cells (HPMECs). These cells were exposed to hyperoxia (0.95 FiO 2 ) for up to 72 hours, and SNAI1 mRNA (Figure 2(a)) and protein expression (Figure 2(b)) were measured. SNAI1 mRNA expression was significantly higher in male HPMECs compared to female. Male HPMECs showed increased SNAIL1 protein upon exposure to hyperoxia, while female HPMECs did not show this change.

3.3.1.
Regulation of miR30a-5p by HIF-1α. We have previously shown that Hif-1α expression was increased in female mice at P7 and decreased in male mice at P21 after hyperoxia exposure [29] and that HIF-1α binding to its target genes is more significant in female lungs after hyperoxia exposure [22]. HIF-1α protein expression was decreased in both male and female HUVECs upon exposure to hyperoxia [29]. Motif analysis revealed a hypoxia response element (HRE) [AG]CGTG site 428 base pairs upstream of miR-30a (Figure 4(a)). To further establish the potential regulation of miR-30a by Hif-1α, we used the siRNA-mediated loss of the Hif-1α expression in female human pulmonary microvascular endothelial cells and its effect on miR30a-5p expression. Treatment of female HPMECs with HIF-1α siRNA significantly decreased HIF-1α expression (Figure 4(b)). Also, the expression of miR30a-5p was decreased in pulmonary microvascular endothelial cells transfected with HIF-1α siRNA compared to cells treated with scrambled siRNA (Figure 4(c)).

SNAI1 Expression following HIF-1α Silencing in Female
Human Pulmonary Microvascular Endothelial Cells. We next wanted to see if HIF-1α played a role in regulating SNAI1 expression in human pulmonary microvascular endothelial cells. Following the silencing of HIF-1α in female HPMECs, SNAI1 mRNA ( Figure 5(a)) and protein expression ( Figure 5(b)) were decreased in female HPMECs.

SNAI1
Expression following miR30a-5p Inhibition, HIF-1α Overexpression, or Both in Female HPMECs. Since both miR30a-5p and HIF-1α had opposing effects on SNAI1 expression, we wanted to determine the independent and additive effects of miR30a-5p inhibition and HIF-1α overexpression in female HPMECs. HIF-1α overexpression and miR30a-5p inhibition both independently increased SNAI1 mRNA expression in female HPMECs (Figure 6(a)).  Figure 2: Snai1 mRNA and protein expression is increased in human pulmonary microvascular endothelial cells upon exposure to hyperoxia: SNAI1 mRNA (a) and protein expression (b) were measured in neonatal human pulmonary microvascular endothelial cells exposed to normoxia or hyperoxia (0.95 FiO 2 up to 72 hrs). Values are mean ± SD. n = 3 per group. Significant differences between indicated groups are indicated by * P < 0:05. HIF-1α overexpression had a more significant effect on SNAI1 expression compared to miR30a-5p inhibition. Combined HIF-1α overexpression and miR30a-5p inhibition increased SNAI1 expression compared to controls, but there was no additive effect ( Figure 6(a)). SNAI1 protein was increased by miR30a-5p inhibition or HIF-1α overexpression independently and by the combined treatment with miR30a-5p inhibition and HIF-1α overexpression compared to controls. No statistically significant changes in protein levels were noted between the individual treatment groups (Figure 6(b)).

Discussion
Sex-specific differences exist in many neonatal morbidities including bronchopulmonary dysplasia (BPD), which adversely affects a significant number of extremely premature newborns. Elucidation of mechanisms responsible for the sex-specific differences in the development of this disease is critical for the development of individualized therapies. We identified miR-30a as a potential modulator of many of the genes differentially expressed in the angiogenesis pathway [9]. Interestingly, miR-30a expression was increased in the female lung after exposure to hyperoxia. These findings were replicated in neonatal human pulmonary microvascular endothelial cells (HPMECs), with higher miR-30a expression in female HPMECs upon exposure to hyperoxia. miR-30a modulates many biological processes and has proangiogenic and antifibrotic properties. A miRNA can have many putative mRNA targets. Snai1 (a profibrotic gene) is a miR-30a target. We show that Snai1 expression is decreased in female neonatal mice in vivo. On the other hand, male HPMECs show increased SNAI1 expression in vitro after exposure to hyperoxia. Also, we show that Hif-1α could modulate miR-30a expression in HPMECs. Interestingly, Hif-1α also increases Snai1 expression, and the net protective effect in females could be mediated through a Hif-1α-mediated increase in the miR-30a expression which keeps the Snai1 expression in check during exposure to hyperoxia (Figure 7). Snai1 (Snai Family Transcriptional Repressor 1) is a zinc finger transcriptional repressor protein and plays a central role in both epithelial to mesenchymal (EMT) [30] and endothelial to mesenchymal (Endo-MT) in many disease processes. Exposure to hyperoxia leads to a profibrotic phenotype leading to an increase in myofibroblasts [31][32][33][34]. Endo-MT is a complex biological process in which endothelial cells lose their surface expression of endothelial-specific markers (e.g., CD31 and vWF) and acquire a mesenchymal phenotype (elongated and fusiform) and express mesenchymal markers (like α-SMA and vimentin) [35]. The role of Endo-MT is known in lung diseases such as idiopathic pulmonary fibrosis (significant contribution to fibrosis and vascular regression) and pulmonary arterial hypertension (by causing pulmonary vascular remodeling and endothelial dysfunction) [36][37][38]. We have previously shown that male HPMECs show higher expression of α-SMA upon exposure to hyperoxia accompanied by a decrease in CD31 expression [39]. In this study, SNAI1 mRNA expression was significantly higher in male HPMECs compared to female. Male HPMECs showed increased SNAIL1 protein upon exposure to hyperoxia, while female HPMECs did not show this change. In vivo, Snai1 expression was decreased in females immediately after (P7) hyperoxia exposure. Since our group had discovered sex-specific differences in miR-30a expression with females showing higher expression than males [9], we wanted to elucidate the possible regulation of SNAI1 expression by miR-30a in female HPMECs. Other studies have shown miR-30a mediated downregulation of Snai1 [10, 14, 17]. Zhang et al. reported that miR-30a targeted and downregulated SNAI1 in a diabetic cataract model in vitro. Similar antifibrotic effects were reported in non-small cell lung cancer [40], atrial fibrillation-induced myocardial fibrosis [14], and TGF-β1-induced peritoneal fibrosis [10]. In female HPMECs, in the present study, we show decreased expression of SNAI1 with miR-30a overexpression and increased expression with miR-30a inhibition. To our knowledge, this is the first study to show that miR-30a targets SNAI1 in neonatal HPMECs.
Even though we had reported sex-specific differences in miR-30a expression [9], upstream modulation of miR-30a was not elucidated. Hif-1α is expressed in early lung development and slowly decreases expression as the lung matures, with a second transient peak at the alveolar stage of lung development around PND10 [20,41]. Use of prolyl hydroxylase domain (PHD) inhibitors to increase HIF protein levels improved lung growth in a preterm baboon model of BPD [42] Interestingly, Hif-1α activity is increased, when human umbilical vein endothelial cells are returned to normoxic conditions after being exposed to hyperoxia [43]. We had previously reported that HIF-1α binding to its target genes is more significant in female lungs after hyperoxia exposure [22]. In contrast to male mice, females did not show a decrease in Hif-1α expression at PND21 after early hyperoxia exposure. Yang et al. showed that HIF-1α inhibition abrogated miR-30a upregulation in cardiomyocytes [18]. In this study, we were able to show that the silencing of HIF-1α in HPMECs significantly decreased miR-30a expression in HPMECs, thus postulating a role for HIF-1α in increasing miR-30a expression in the hyperoxia model, especially during the normoxic recovery phase after exposure to hyperoxia.
Many factors including HIF-1α, NFKB, STAT3, and NICD can induce Snai1 expression [30]. Xu et al. showed that HIF-1α could induce Snai1, independent of TGF-β, and that SNAI1 is a direct target of HIF-1α in human coronary endothelial cells [23]. We wanted to see if HIF-1α modulated SNAI1 expression in HPMECs. HIF-1α knockdown in HPMECs decreased SNAI1 mRNA and protein expression. Similar results were reported by Zhu et al. in pancreatic cancer cells [44]. Hypoxia-induced Endo-MT in a model of radiation-induced pulmonary fibrosis was mediated through HIF-1α [45]. HIF-1α may be thus inducing both profibrotic (SNAI1) and antifibrotic (miR-30a) pathways in HPMECs. The balance of expression between the two mediators may eventually predispose or protect the lung from pathologic fibrosis. In our study, HIF-1α overexpression and miR30a-5p inhibition either separately or in combination increased SNAI1 expression, but the combined treatment did not produce an additive effect in SNAI1 expression.
We recognize the following limitations in this study. Our study only evaluated two time points in vivo after hyperoxia exposure during the saccular stage of lung development in mice. Later time points may have sustained differences in Snai1 expression, or these differences may be lost. Also, we used high oxygen concentration (0.95 FiO 2 ) for our hyperoxia exposure protocol based on the reproducible BPD-like phenotype produced in mice. Even though this limits hyperoxia exposure to the saccular stage of lung development, which equates to 26-36 weeks in human neonates, this oxygen concentration is much higher than what is used in the human NICU. Also, these results may be different from other models, where hyperoxia exposure is continued for longer durations [2]. We are also aware that findings in the murine model may be strain-specific [46]. Our in vivo data is from whole lung tissue and does not show endothelial-specific expression. Our future studies will seek to validate the cellspecific expression of Snai1 in primary mouse endothelial cells and BPD and non-BPD patient samples. In conclusion, we show that the intersection between HIF-1α, miR-30a, and Snai1 may play a role in driving sex-specific differences in after neonatal hyperoxic lung injury.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper.